Can AI save the planet?

Generative Artificial Intelligence (GenAI) promises to be more transformative than any other technology in recent memory, with the power to give rise to new industries and professions while propelling the global economy toward a new era of prosperity. Demand for data processing is expected to grow exponentially with AI, as data is essential to how AI systems learn and make decisions. AI’s ability to turn mountains of data into insights requires compute power, making data centre performance critical to our ability to leverage this transformative technology.

As with any generational technology, there will be challenges. A key – but solvable – hurdle is the anticipated environmental cost of AI due to the energy and data centre resources required to run larger computing models. Organisations seek to embrace the many benefits of AI while also making progress toward their sustainability goals.

Data centre energy use and emissions are serious issues the industry is addressing head on. When approached mindfully, AI infrastructure development and adoption in data centres can provide a path to more sustainable operations. As experts in data and IT infrastructure, at Dell Technologies we believe sustainability will be integral to the success of AI technology, in South Africa and across the continent. We have the end-to-end solutions and our own company commitment to not only help offset the environmental impact of AI, but to also harness the tremendous potential for AI to support climate-related solutions.

In May 2023, Gartner data supported the business community’s ongoing commitment to sustainability noting that in FY23, “most CEOs (94%) will increase or hold sustainability and ESG investments at similar levels to 2022”. Then GenAI came roaring onto the scene, shifting the focus to this burgeoning technology. In fact, our research found that currently, 76% of IT decisionmakers plan to increase their budgets to support GenAI use cases and 78% are excited about how an investment in AI can benefit their organisations.

At first glance, these findings may appear to be at odds with one another, as companies are simultaneously increasing investment in sustainability and in energy-intensive technology. We believe sustainability and digital transformation is not an ‘either/or’ decision. In fact, technological progress is a prerequisite for companies seeking to meet ambitious climate goals. We are tech optimists – and we’re also pragmatic. We understand the best innovations not only advance our technological capacity and allow us to optimise data, but do so while supporting more energy efficient and sustainable futures.

_{Doug Woolley, GM of Dell Technologies South Africa.}

While AI requires significant compute power, it currently represents a small fraction of IT’s global energy consumption. We expect this will change as more companies, governments and organisations harness AI to drive efficiency across their operations. To manage AI’s growing carbon footprint, data centre operators must embrace sustainable data centre investments and practices. Data centre energy use remained stable over the past decade (around 1% of global electricity demand according to the International Energy Agency), even with growing technology workloads and traffic.

A recent study by the South African Data Centre Association indicates that data centres in South Africa consumed approximately 2.37 TWh of electricity in 2022, accounting for nearly 1.3% of the country’s total electricity consumption. This raises concerns about the environmental impact as well as the strain on South Africa’s struggling power infrastructure.

To offset the environmental impact of AI, greater control over data centre energy consumption is increasingly becoming a top priority, and there are tools available to do this:

• Minimise AI’s carbon footprint through modern, energy-efficient servers and storage devices as well as environmentally responsible cooling methods while powering data centres with renewable energy. At Dell, we prioritise running larger data models in our data centres that are powered by 100% renewable energy.
• Right-size AI workloads and data centre economics. While some organisations will benefit from general purpose large language models (LLMs), many organisations only require domain- or enterprise-specific implementations. Right-sizing compute requirements and infrastructure can support greater data centre efficiency. And, flexible ‘pay as you go’ spending models can also help organisations save on data centre costs while supporting sustainable IT infrastructure.
• Responsibly retire inefficient hardware to optimise data centre performance and energy consumption while reducing e-waste and keeping recycled materials in use longer.  

AI solutions to environmental challenges

As sustainable data centres can help to offset AI’s carbon footprint, this technology can also be used to track and analyse massive amounts of data to ultimately address some of our planet’s biggest challenges, such as climate change, pollution and deforestation. For example, Dell is helping Siemens build smarter buildings with AI. Siemens helps customers reduce their buildings’ carbon footprints by leveraging edge and AI technologies to address building performance issues, like optimising HVAC systems, predicting energy demand and identifying energy leaks in real time.

AI can be used to optimise energy grids, design more efficient transportation systems and develop new ways to capture and store carbon dioxide. The University of Cambridge and Dell collaborated to support advanced and sustainable research through AI. By providing powerful, energy-efficient supercomputing to scientists and organisations, the University of Cambridge drives breakthroughs in innovation that rely on AI. This solution not only enables AI to process enormous volumes of data more quickly, but it does so more efficiently, with less power consumed.

Within data centre operations, AI can be used to improve monitoring and workload placement to optimise efficiency and reduce energy costs. There is no ‘either/or’ decision; rather, efficient data centre infrastructure is integral to AI’s evolution. For this to succeed, advocates and organisations must see sustainability as a vital part of AI computing infrastructure. According to IDC, the number one sustainability priority for IT planning and procurement among IT decision makers is reducing data centre energy consumption ². At the intersection of sustainability and business priorities, AI can support environmental stewardship at the same time it drives digital transformation. 

Leading by example

Technology has an important role in addressing environmental challenges. In fact, according to a Dell Technologies study, almost two-thirds (64%) of Gen Z believe technology will play an important role in the fight against the climate crisis. Dell aims to move the industry forward both through modernising data centre technology and modelling the ‘both/and’ benefits of sustainable data centres. We have ambitious goals and we also see the immense benefits AI can bring to protecting our planet. While working to offset the environmental impact of AI, we will also innovate to develop solutions that leverage the power of AI to address some of our biggest environmental challenges. To learn more, visit Artificial Intelligence.